Skip to content

CI: Canary-test both zero-shot and chain-of-thought #4

CI: Canary-test both zero-shot and chain-of-thought

CI: Canary-test both zero-shot and chain-of-thought #4

name: General Knowledge
on: [workflow_dispatch, push, pull_request]
jobs:
general-knowledge:
runs-on: ubuntu-22.04
timeout-minutes: 10
strategy:
max-parallel: 3
fail-fast: false
matrix:
model:
- mistralai/mistral-7b-instruct-v0.3 # $0.07 [32K]
- meta-llama/llama-3-8b-instruct # $0.07 [ 8K]
- openchat/openchat-7b # $0.07 [ 8K]
- openchat/openchat-8b # $0.08 [ 8K]
- qwen/qwen-4b-chat # $0.09 [32K]
- microsoft/phi-3-medium-4k-instruct # $0.14 [ 4K]
- nousresearch/hermes-2-pro-llama-3-8b # $0.15 [ 8K]
- teknium/openhermes-2.5-mistral-7b # $0.17 [ 4K]
- nousresearch/nous-hermes-2-mistral-7b-dpo # $0.18 [32K]
- open-orca/mistral-7b-openorca # $0.18 [ 8K]
- mistralai/mixtral-8x7b-instruct # $0.24 [32K]
- qwen/qwen-14b-chat # $0.27 [32K]
- meta-llama/llama-3-70b-instruct # $0.59 [ 8K]
- mistralai/mixtral-8x22b-instruct # $0.65 [64K]
steps:
- uses: actions/checkout@v4
- run: ./query-llm.js tests/canary-single-turn.txt
env:
LLM_API_BASE_URL: ${{ secrets.LLM_API_BASE_URL }}
LLM_API_KEY: ${{ secrets.LLM_API_KEY }}
LLM_CHAT_MODEL: ${{ matrix.model }}
- run: ./query-llm.js tests/general-knowledge.txt
env:
LLM_API_BASE_URL: ${{ secrets.LLM_API_BASE_URL }}
LLM_API_KEY: ${{ secrets.LLM_API_KEY }}
LLM_CHAT_MODEL: ${{ matrix.model }}